Influence of soil properties, topography, and land cover on soil organic carbon and total nitrogen concentration: A case study in Qinghai-Tibet plateau based on random forest regression and structural equation modeling

The effects of topography, land cover type, and soil physicochemical properties on the distribution of soil organic carbon (SOC) and total nitrogen (TN) have drawn growing attention recently, but little is known about how these features-associated interactions impact SOC and TN. To elucidate how these interactions affect the preservation of carbon and nitrogen in soils, we used data-driven models (random forest regression and structural equation modeling) to identify the dominant environmental factors affecting the distribution of SOC and TN in two different soil layers (0-20 and 20-40 cm) of the Qinghai-Tibet plateau. In addition, an algorithm based on random forest ("Boruta") was chosen to identify the relevant influencing factors and partial dependence was used to depict the two most important factors. We found that rather than land cover type, environmental properties, such as soil physicochemical characteristics and altitude had the most significant effects on the distribution of SOC and TN. Our findings indicate that elevation and TN are the two most important factors influencing SOC in the surface and subsurface soil layers. Moreover, total potassium (TK) impacts TN content in the surface soil layer, but only in a specific range of concentrations, which could be attributed to anthropogenic activities such as applying nitrogen and potassium fertilizers to increase the yields of local food crop, Tibetan hulless barley. These findings provide a scientific perspective on soil nutrient preservation.